GASTROENTEROLOGY 2000;118:735–748
`
`Molecular Mechanism of Interferon Alfa–Mediated Growth
`Inhibition in Human Neuroendocrine Tumor Cells
`
`KATHARINA M. DETJEN,* MARTINA WELZEL,* KATRIN FARWIG,* FELIX H. BREMBECK,‡
`ASTRID KAISER,‡ ERNST–OTTO RIECKEN,‡ BERTRAM WIEDENMANN,* and STEFAN ROSEWICZ*
`*Medizinische Klinik mit Schwerpunkt Hepatologie und Gastroenterologie, Universita¨tsklinikum Charite´, Campus Virchow Klinikum, Humboldt
`Universita¨t zu Berlin; and ‡Innere Medizin I, Gastroenterologie und Infektiologie, Universita¨tsklinikum Benjamin Franklin, Freie Universita¨t Berlin,
`Berlin, Germany
`
`Background & Aims: Although human neuroendocrine
`tumors respond to interferon (IFN)-a treatment in vivo,
`the underlying mechanisms of growth inhibition are
`poorly understood. To characterize the antiproliferative
`effects at a molecular level, we explored the growth-
`regulatory action of IFN-a in the human neuroendo-
`crine tumor cell lines BON and QGP1. Methods: IFN-a
`receptor expression and signal transduction were exam-
`ined by reverse-transcription polymerase chain reac-
`tion, immunoblotting, subcellular fractionation, and
`transactivation assays. Growth regulation was evalu-
`ated by cell numbers, soft agar assays, and cell cycle
`analysis using flow cytometry. Expression and activity
`of cell cycle–regulatory molecules were determined by
`immunoblotting and histone H1–kinase assays. Results:
`Both cell lines expressed IFN-a receptor mRNA tran-
`scripts. Ligand binding initiated phosphorylation of Jak
`kinases and Stat transcription factors, resulting in Stat
`activation, nuclear translocation, and transcription from
`an ISRE-reporter construct. Prolonged IFN-a treatment
`dose-dependently inhibited both anchorage-depen-
`dent and -independent growth. Cell cycle analysis of
`IFN-a–treated, unsynchronized cultures revealed an
`increased S-phase population, which was further sub-
`IFN-a–
`stantiated in G1 synchronized QGP1 cells.
`treated cells entered S phase in parallel to control
`cultures, but their progress into G2/M phase was
`delayed. Both cellular cyclin B levels and CDC 2 activity
`were substantially reduced. The extent and time course
`of this reduction corresponded to the observed S-
`phase accumulation. Conclusions: IFN-a directly inhib-
`its growth of human neuroendocrine tumor cells by
`specifically delaying progression through S phase and
`into G2/M. These cell cycle changes are associated with
`inhibition of cyclin B expression, resulting in reduced
`CDC2 activity.
`
`Neuroendocrine (NE) gastroenteropancreatic (GEP)
`
`tumors constitute a biologically heterogeneous group
`of rare, predominantly slow-progressing but mostly
`malignant neoplasms.1 They frequently synthesize and
`
`secrete bioactive molecules, such as regulatory peptides,
`kinins, and serotonin, which are responsible for the
`tumor-associated hypersecretion syndrome.2 Because more
`than 80% of malignant GEP tumors present liver
`metastases at the time of diagnosis, systemic therapy is
`usually required.1
`Current therapeutic strategies aim to control both
`hypersecretion-related symptoms and tumor growth.3–5
`Somatostatin and its analogues are effective in suppress-
`ing the hypersecretion syndrome, but they are frequently
`insufficient to control tumor progression.6 Therefore, multi-
`modal therapeutic approaches are evaluated that encom-
`pass surgical debulking, chemoembolization, chemother-
`apy, and biotherapy.3,5,7 The typically low proliferative
`fraction of NE tumors limits the efficacy of conventional
`chemotherapeutic approaches, so that conclusive prolon-
`gation of survival remains to be demonstrated.8,9 In view
`of the potential side effects of chemotherapeutic regimes,
`biotherapeutic approaches using long-term treatment
`with interferon (IFN)-a alone or in combination with soma-
`tostatin analogues have widely replaced conventional
`chemotherapeutic strategies over the past decade.3–5,7,10
`Patients with advanced GEP NE tumor disease have
`received IFN-a in the context of clinical trials; a
`significant fraction of tumors responded to IFN-a treat-
`ment, as determined by biochemical markers such as
`5-hydroxyindolacetic acid excretion or chromogranin A
`serum levels.5,7,11 IFN-a–induced reduction of such
`biochemical markers was reflected by a marked relief of
`hypersecretion-associated symptoms in functionally ac-
`tive tumors. Several of these studies also reported an
`
`Abbreviations used in this paper: DMEM, Dulbecco’s modified
`Eagle medium; DTT, dithiothreitol; FCS, fetal calf serum; GED,
`gastroenteropancreatic; IFN, interferon; ISRE, interferon-stimulated
`response element; PKR, RNA-dependent protein kinase; PMSF,
`phenylmethylsulfonyl fluoride; SDS-PAGE, sodium dodecyl sulfate–
`polyacrylamide gel electrophoresis.
`r 2000 by the American Gastroenterological Association
`0016-5085/00/$10.00
`doi:10.1053/gg.2000.5968
`
`Roxane Labs., Inc.
`Exhibit 1039
`Page 001
`
`

`
`736 DETJEN ET AL.
`
`GASTROENTEROLOGY Vol. 118, No. 4
`
`IFN-a–mediated inhibition of tumor progression result-
`ing in stable disease or, in some cases, partial remission
`with reduction of tumor burden.11,12 It is of particular
`clinical relevance that inclusion of IFN-a in the therapeu-
`tic regimen enabled suppression of disease progression in
`a significant fraction of tumors resistant to monotherapy
`with somatostatin analogues.5 Nonetheless, a significant
`fraction of endocrine GEP tumors fail to respond to the
`growth-inhibitory effects of IFN-a. The clinical benefit
`of the specific antiproliferative actions of IFN-a has
`therefore remained controversial.
`Multiple mechanisms have been proposed to contrib-
`ute to overall clinical outcome of IFN-a treatment,
`including direct antiproliferative actions of IFN-a on the
`tumor cells and indirect mechanisms, such as inhibition
`of angiogenesis, induction of tumor mesenchyme, and
`immunomodulation via up-regulation of major histocom-
`patibility class I antigens (reviewed by Grander et al.13).
`Resistance of
`individual NE tumors to the growth-
`regulatory effects of IFN-a may result from defective
`signal transduction or tumor specific phenotypic alter-
`ations in growth-regulatory pathways. Because the respon-
`siveness of the secretory pathway is frequently main-
`tained in functionally active GEP tumors that are
`refractory to the growth-inhibitory actions of IFN-a,10–12
`the latter mechanism seems to be of major biological
`relevance in endocrine GEP tumors. Consequently, iden-
`tification of the specific growth-relevant cellular targets
`of IFN-a and understanding of their function are needed
`to maximally exploit the potential of IFN-a in the
`treatment of NE tumor disease.
`Much progress has been made in elucidating the signal
`transduction pathways activated by type I IFNs.14,15
`Ligand binding to specific receptors at the plasma
`membrane results in dimerization of receptor subunits
`and thereby initiates the rapid autophosphorylation of
`receptor-associated Janus tyrosine kinases ( JAKs) Jak1
`and Jak2. The activated JAKs
`in turn tyrosine-
`phosphorylate and activate latent cytosolic members of
`the Stat transcription factor family, named for their dual
`function as signal transducers and activators of transcrip-
`tion. Activated Stat1 and Stat2 transcription factors then
`associate with a 48-kilodalton protein into multimeric
`complexes termed IFN-stimulated gene factor 3 (ISGF3)
`that translocate into the nucleus where they induce
`changes in gene expression, which ultimately result in
`the biological effects of IFN-a treatment.
`In contrast to these well-established and conserved
`signal transduction events, the molecular basis of the
`antiproliferative effects of IFN-a appears to be highly cell
`type specific and has not yet been delineated in GEP NE
`tumor cells. The most comprehensive analysis of IFN-a–
`
`induced growth regulation used Daudi Burkitt lym-
`phoma cells and identified multiple signaling pathways
`that apparently function in parallel to block G1–S
`progression. Two major independent and complementary
`effector pathways have emerged from these studies. The
`first is accumulation of the hypophosphorylated, growth-
`restrictive form of the retinoblastoma-associated tumor-
`suppressor protein Rb.16–21 This functional activation of
`Rb is attributed to down-regulation of G1 cyclins and
`CDC25 phosphatases in conjunction with induction of
`cyclin-dependent kinase inhibitors (CKIs) and total Rb
`content. The second pathway is induction of the double-
`stranded RNA–activated protein kinase (p68 PKR)
`initiating down-regulation of cellular c-myc levels.22,23
`However, alternative mechanisms may prevail in NE
`tumors that have been reported to exhibit functional
`inactivation of Rb24 and overexpression of c-myc.25 To
`specifically evaluate the direct cell cycle–regulatory ef-
`fects of IFN-a on GEP NE tumor cells, we analyzed
`IFN-a–dependent signal transduction and growth regu-
`lation in an in vitro model of human GEP NE tumor
`disease.
`Materials and Methods
`Materials
`QGP1 cells were kindly provided by K. Mo¨lling
`(Zu¨ rich, Switzerland); BON cells were a generous gift from
`C. M. Townsend (Galveston, TX). Dulbecco’s modified Eagle
`medium (DMEM), RPMI 1640 medium, and phosphate-
`buffered saline (PBS) were supplied by GIBCO BRL (Berlin,
`Germany). Fetal calf serum (FCS), trypsin/EDTA, penicillin,
`and streptomycin were from Biochrom (Berlin, Germany). The
`antibodies were purchased from the following manufacturers:
`Santa Cruz Biochemicals (Santa Cruz, CA; CDK2, Jak1, Tyk2,
`Stat1, Stat2); Transduction Laboratories
`(Lexington, KY;
`p27kip1); Pharmingen (San Diego, CA; cyclins E, A, and B);
`Calbiochem–Novabiochem GmbH (Bad Soden, Germany;
`p21cip1); and Dianova GmbH (Hamburg, Germany; all second-
`ary antibodies). [g-32P]Adenosine triphosphate (ATP) and
`enhanced chemiluminescence Western blotting detection re-
`agents were obtained from Amersham (Braunschweig, Ger-
`many), and calf histone H1, deoxyribonuclease, and ATP from
`Boehringer Mannheim (Mannheim, Germany). Polymerase
`chain reaction (PCR) reagents were obtained from Promega
`(Heidelberg, Germany), and Western blot supplies from
`Bio-Rad Laboratories GmbH (Mu¨ nchen, Germany). IFN-a2b
`(Roferon) was kindly provided by Hofmann-LaRoche (Basel,
`Switzerland). Protein A–Sepharose beads and all other reagents
`were purchased from Sigma Chemical Co. (Deisenhofen, Ger-
`many).
`
`Cell Lines and Tissue Culture
`The human NE pancreatic tumor cell lines QGP1 and
`BON were grown as subconfluent monolayer cultures in RPMI
`
`Roxane Labs., Inc.
`Exhibit 1039
`Page 002
`
`

`
`April 2000
`
`IFN–a IN NEUROENDOCRINE TUMOR CELLS 737
`
`1640 and DMEM, respectively, supplemented with 10%
`(vol/vol) FCS, 100 U/mL penicillin, and 100 µg/mL streptomy-
`cin, and kept in 95% air and 5% CO2 at 37°C. Experiments
`were routinely performed in the log phase of growth after cells
`had been plated for 16–24 hours. For experiments on synchro-
`nized cultures, QGP1 cells were serum starved for 36 hours,
`which resulted in G1-phase accumulation of approximately
`70% of the cells, and were then restimulated by addition of
`10% FCS.
`
`Cell Growth Assays
`Effects of IFN-a on proliferation of NE tumor cells
`were evaluated by determination of cell numbers. Cells were
`plated at 50,000 cells/well (BON) or 25,000 cells/well (QGP1)
`in 12-well tissue culture plates, and cells were manually
`counted using a hemacytometer at indicated time points.
`Viability of cells was confirmed by trypan blue exclusion and
`was routinely .95%. To determine effects of IFN-a on
`anchorage-independent growth of NE tumor cells, colony
`formation in agar suspension was evaluated. Briefly, 3 3 104
`cells were resuspended in 300 µL of culture medium and added
`to a mixture of 2.7 mL Hyclone FCS, 0.8 mL Iscove’s modified
`DMEM, 3.6 mL of 2.1% (wt/vol) methylcellulose in Iscove’s,
`1.6 mL agar solution (10 mL 3% agar and 20 mL DMEM), and
`0.06 mL b-mercaptoethanol to obtain an agar suspension.
`Aliquots (1 mL, 3000 cells) of this suspension were then
`dispensed in 35-mm dishes containing the indicated concentra-
`tions of IFN-a or vehicle. Colony formation was assessed under
`an inverted microscope by manual counting after a 10-day
`incubation period. A threshold of 20 cells was arbitrarily set to
`score cell accumulations as colonies.
`
`Flow-Cytometric Analysis
`For cell cycle analysis, approximately 106 cells were
`harvested by gentle trypsinization (0.25%), carefully resus-
`pended in PBS, and fixed in ethanol (70%) for 30 minutes at
`–20°C. After brief centrifugation, cells were washed once with
`PBS and incubated for 30 minutes at room temperature in PBS
`containing 100 µg/mL ribonuclease A, 0.1 % Triton X-100, 1
`µmol/L EDTA, and 1.5 µg/mL propidium iodide. Cell cycle
`analysis was carried out on a FACScan using Modfit software
`(Becton Dickinson, Heidelberg, Germany).
`
`Western Blotting and Subcellular
`Fractionation
`For analysis of cell cycle proteins, subconfluent cells
`were treated as indicated, rinsed twice with ice-cold PBS
`containing 1 mmol/L sodium orthovanadate (Na3VO4), and
`extracted in ice-cold lysis buffer (20 mmol/L Tris [pH 7.8], 150
`mmol/L NaCl, 2 mmol/L EDTA, 50 mmol/L b-glycerophos-
`phate, 0.5% Nonidet P-40, 1% glycerine, 10 mmol/L NaF, 1
`mmol/L sodium orthovanadate, 1 mmol/L dithiothreitol [DTT],
`2 µmol/L phenylmethylsulfonyl fluoride [PMSF], 10 µg/mL
`aprotinin, and 2 µmol/L leupeptin). Extracts were boiled in
`Laemmli’s sample buffer, and aliquots were separated by
`sodium dodecyl sulfate–polyacrylamide gel electrophoresis
`
`(SDS-PAGE), electroblotted to polyvinyl difluoride mem-
`branes (New England Nuclear, Ko¨ln, Germany), and blocked
`in PBS/0.1% Tween (PBST) containing 5% nonfat dry milk
`(cyclins, E, A, B, p21cip1, p27kip1, Rb, p16, Jak1, Tyk2, Stat1,
`Stat2) or 5% bovine serum albumin (CDK2, CDC2) for 2 hours
`at room temperature. Incubation with primary antibodies was
`carried out overnight at 4°C and was followed by 3 washes in
`PBST at room temperature. After incubation with appropriate
`horseradish peroxidase–conjugated secondary antibodies, mem-
`branes were washed 3 times for 15 minutes in PBST, and bands
`were visualized by enhanced chemiluminescence following the
`manufacturer’s recommendations. Results were quantitated by
`laser densitometry using the Scanpak 2 software (Biometra,
`Go¨ttingen, Germany).
`For Stat translocation studies, subcellular fractions of QGP
`cells were prepared before immunoblotting according to the
`procedure described by Simboli-Campbell.26 After hypotone
`lysis (1 mmol/L NaHCO3, 5 mmol/L MgCl2, 0.1 mmol/L
`PMSF, and 20 µg/mL leupeptin), extracts were adjusted to 50
`mmol/L Tris-HCl (pH 7.5) plus 50 mmol/L EGTA and nuclei
`were isolated by brief centrifugation at 500g. Nuclei were
`further purified by saccharose gradient centrifugation (45%
`[wt/vol] saccharose, 30 minutes, 1660g) and then lysed in
`buffer A containing 50 mmol/L Tris-HCl
`(pH 7.5), 0.5
`mmol/L EGTA, 0.5 mmol/L EDTA, 1% b-mercaptoethanol, 1
`mmol/L PMSF, 20 µg/mL leupeptin, and 1% SDS by brief
`sonication. Aliquots of the nuclear and the cytosolic fractions
`were then processed as described above.
`
`Immunoprecipitation and Histone H1
`–Kinase Assays
`To determine tyrosine phosphorylation of Jak kinases
`and Stat proteins, subconfluent cells were treated as indicated,
`then washed twice in ice-cold PBS and lysed by brief sonication
`in IP buffer (20 mmol/L Tris [pH 7.8], 150 mmol/L NaCl, 2
`mmol/L EDTA, 0.5% Nonidet P-40, 10 mmol/L NaF, 10
`mmol/L sodium orthovanadate, 2 mmol/L PMSF, and 10
`µg/mL each of leupeptin, pepstatin, and aprotinin). The lysates
`were precleared for 1 hour by incubation with protein A–Sepha-
`rose beads. Immune complexes were collected on protein
`A–Sepharose beads that had been coated with saturating
`amounts of primary antibody. Beads were subsequently washed
`4 times with ice-cold IP buffer, boiled in Laemmli’s sample
`buffer, separated by SDS-PAGE, and transferred to polyvinyl
`difluoride membranes. Nonspecific binding was blocked with
`10 mmol/L Tris-HCl (pH 7.6), 150 mmol/L NaCl, and 0.05%
`Tween 20 containing bovine serum albumin (1%) for 2 hours at
`room temperature. The membrane was then incubated with
`a-phosphotyrosine antibody (1 µg/mL) for 2 hours at room
`temperature and further processed as described for Western
`blots. To control that equivalent amounts of immunocomplexes
`were analyzed, membranes were reprobed with the antibodies
`used for immunoprecipitation. For repeated use, membranes
`were stripped by incubating the membrane in a solution of
`62.5 mmol/L Tris-HCl (pH 6.8), 2% SDS, and 100 mmol/L
`2-mercaptoethanol for 30 minutes at 55°C followed by a
`
`Roxane Labs., Inc.
`Exhibit 1039
`Page 003
`
`

`
`738 DETJEN ET AL.
`
`GASTROENTEROLOGY Vol. 118, No. 4
`
`washing in PBST for 2 hours, changing the buffer every 30
`minutes.
`To examine the composition and activity of cdk complexes,
`immunoprecipitations were performed. For immunoprecipita-
`tion of CDK2 or CDC2, cells were lysed by mild sonication in
`ice-cold ELB buffer (50 mmol/L HEPES [pH 7.5], 250 mmol/L
`NaCl, 5 mmol/L EDTA, 0.1% Nonidet P-40, 1 mmol/L DTT,
`1 mmol/L NaF, 0.1 mmol/L Na3VO4, 2 µg/mL aprotinin, 5
`µg/mL leupeptin, and 0.1 mmol/L PMSF)
`followed by a
`30-minute incubation on ice with occasional vortexing. After
`brief centrifugation (10 minutes, 4°C, 13,000g), lysates (500
`µg/sample) were precleared for 1 hour by incubation with
`protein A–Sepharose beads. Immuncomplexes were then pre-
`cipitated by addition of protein A–Sepharose beads precoated
`with saturating amounts of CDK2 or CDC2 antibody (10 µg).
`Samples were incubated with gentle agitation at 4°C for 4
`hours. Immuncomplexes were subsequently washed 4 times in
`ice-cold ELB buffer and twice in 50 mmol/L HEPES (pH 7.5)
`containing 1 mmol/L DTT. The kinase reaction was then
`started by addition of 30 µL kinase buffer containing 50
`mmol/L HEPES (pH 7.5), 1 mmol/L DTT, 10 mmol/L MgCl2,
`1 µg calf histone H1 per sample, 50 µmol/L ATP, and 5 µCi
`[g-32P]ATP per sample. The kinase reactions were allowed to
`proceed for 5 minutes (CDK2) and 30 minutes (CDC2) and
`were then terminated by boiling the samples in Laemmli’s
`buffer. The whole reaction was subjected to 10% SDS-PAGE,
`and kinase activities were determined by autoradiography of
`the dried gels.
`
`Transactivation Studies
`To determine transactivation after IFN-a stimulation
`in BON NE tumor cells, transient transfection assays with an
`IFN-stimulated response element (ISRE)-luciferase reporter
`gene construct were performed. The ISRE-luc construct,
`containing residues 2206/286 of the human 28,58-oligoA
`synthetase enhancer, and a mutated variant were a generous gift
`from Shoumo Bhattacharya27 and were subcloned into pTGL2
`promoter vector (Promega, Madison, WI). BON cells (5 3 105
`cells/well) were plated in 6-well tissue culture dishes, and
`transfections were carried out by calcium phosphate precipita-
`tion technique using a DNA transfection kit (5 Prime-3 Prime;
`Boulder, CO), exactly as described previously.28 Transfected
`cells were allowed to recover for 16 hours and then stimulated
`for 24 hours with the indicated IFN-a concentrations. After
`lysis of cells, the luciferase activity was determined using
`luciferin, ATP, and coenzyme A (Promega) exactly as reported
`previously.28
`
`RNA Preparation and RT-PCR
`To confirm the presence of IFN-a receptor messenger
`RNA (mRNA) transcripts in QGP1 and BON cells, IFN-a
`receptor mRNA was amplified by RT-PCR. Total RNA from
`NE tumor cells was prepared using RNAzol R reagent (WAK
`Chemie, Bad Soden, Germany) according to the manufacturer’s
`instructions. RNA was submitted to deoxyribonuclease diges-
`tion, and aliquots of 1 µg were used for reverse transcription
`
`using Moloney murine leukemia virus reverse transcriptase.
`The reaction was diluted 1:25 for use in the subsequent PCR
`reaction with sequence-specific primers to a 639 base pair
`(bp)-spanning region of the human IFN-a receptor complemen-
`tary DNA (cDNA) (58-AGC GAT GAG TCT GTC GGG;
`38-GGC GTG GAG CCA CTG AAC). Amplification condi-
`tions were exactly as previously described.29
`Results
`Human NE Tumor Cells Express Functional
`IFN-a Receptors
`As a first step to establish the human endocrine
`GEP tumor cell lines QGP1 and BON as a suitable in
`vitro model to investigate IFN-a actions, we determined
`expression of IFN-a receptors at the mRNA level by
`RT-PCR analysis. Using sequence-specific primers against
`the IFN-a receptor cDNA, amplificates of the predicted
`size (639 bp) were detected in both NE cell lines as well
`as in Capan1 exocrine pancreatic cancer cells, which were
`included as positive control29 (Figure 1).
`To confirm that the mRNA transcripts corresponded
`to functionally active receptors, ligand-initiated signal
`transduction was examined next. IFN-a receptors consist
`of 2 chains that associate with Jak1 and Tyk2 kinases.
`Upon IFN-a binding to the receptor, Jak1 and Tyk2
`kinases approach each other and become activated by
`phosphorylation on tyrosine residues. To investigate
`IFN-a–dependent phosphorylation of JAKs in NE tumor
`cell lines, Jak1 (Figure 2, upper panel) and Tyk2 (Figure
`2, lower panel) were immunoprecipitated from whole-
`cell lysates and subsequently analyzed for tyrosine pho-
`phorylation by immunoblotting. After IFN-a stimula-
`tion, Jak1 immunoprecipitates from QGP 1 cells readily
`increased their complement of phosphotyrosine, starting
`as early as 1 minute and reaching a maximum at 15
`minutes. In contrast to the results obtained in QGP1
`
`Figure 1. NE tumor cells express IFN-a receptor mRNA. Total RNA was
`extracted from QGP1, BON, and Capan1 cell
`lines, reverse-tran-
`scribed, and amplified by PCR using primers directed against the type I
`IFN receptor a chain. For size determination of the obtained PCR
`fragment, a 100-bp DNA standard was used (lane 1). Alternating
`sample lanes represent reactions with (1) or without (2) addition of
`reverse transcriptase.
`
`Roxane Labs., Inc.
`Exhibit 1039
`Page 004
`
`

`
`April 2000
`
`IFN–a IN NEUROENDOCRINE TUMOR CELLS 739
`
`Figure 2. IFN-a treatment initiates activation of receptor-associated kinases in human NE tumor cells. Immunoblots of Jak1 and Tyk2
`immunoprecipitates analyzing IFN-a–induced changes in the phosphotyrosine content of IFN-a receptor–associated kinases in (A) QGP1 and (B)
`BON cells. Cells were incubated with 1000 IU/mL IFN-a for the indicated periods. Blots were subsequently stripped and reprobed with antibodies
`against Jak1 or Tyk2 to confirm that equal amounts of immunocomplexes were evaluated. Because Jak1 phosphotyrosine signals were notably
`weaker in BON cells than in QGP1 cells, the autoradiography had to be exposed for 30 minutes compared with 1–2-minute exposures required in
`QGP1 cells.
`
`cells, little increase in Jak1 tyrosine phosphorylation was
`observed in IFN-a–treated BON cells (Figure 2B),
`although weak tyrosine phosphorylation was consistently
`observed under control conditions. However, both QGP1
`and BON cells demonstrated a distinct, time-dependent
`increase in Tyk2 tyrosine phosphorylation, indicative of
`Tyk2 activation at 1, 5, and 15 minutes after IFN-a
`stimulation. Subsequent immunoblotting of the stripped
`phosphotyrosine blots with antibodies against Jak1 (Fig-
`ure 2, upper panel, second row) and Tyk2 (Figure 2, lower
`panel, second row) confirmed that equal amounts of
`immunocomplexes had been analyzed in IFN-a–treated
`cells and control cultures.
`In response to ligand-dependent IFN-a receptor activa-
`tion, latent cytoplasmic Stat1 and Stat2 transcription
`factors are recruited to specific binding sites on the
`cytoplasmic portion of the receptor, where Jak kinases
`then activate these Stat proteins via phosphorylation on
`tyrosine residues. Tyrosine phosphorylation of Stat pro-
`teins was therefore examined to confirm IFN-a–induced
`
`activation of downstream effector molecules. In QGP1
`(Figure 3A) and BON cells (Figure 3B), a pronounced,
`time-dependent increase in Stat 1 tyrosine phosphoryla-
`tion (Figure 3, upper panel) was evident. Enhanced
`phosphorylation was detected on both the 91- and
`84-kilodalton splice variants, with stronger signals for
`the 91-kilodalton form most likely reflecting its higher
`abundance in human NE cells (Figure 3, upper panel,
`second row). Similarly, phosphorylation of Stat2 was
`enhanced (Figure 3, lower panel), although the increase
`was moderate compared with the changes observed for
`Stat1. Comparison of the time courses of Stat activation
`in both cell lines revealed slightly different kinetics.
`Phosphorylation of Stat proteins in QGP1 was transient,
`with maximal effects occurring at 15 minutes of IFN-a
`stimulation for Stat1 and at 5 minutes for Stat2. In BON
`cells, phosphorylation of Stat1 and Stat2 was first evident
`at 5 minutes and thereafter persisted to 60 minutes with
`Stat1 phosphotyrosine levels showing a continuous in-
`crease over this period of time.
`
`Figure 3. IFN-a time-dependently induces tyrosine phosphorylation of Stat1 and Stat2 in NE tumor cells. Both tyrosine phosphorylated (upper
`row) and total (lowerrow). Stat complement of Stat1 (upperpanel) and Stat2 immunoprecipitates (lowerpanel) from IFN-a–treated cells were
`evaluated by immunoblotting. Cells were treated with 1000 IU/mL IFN-a for the indicated intervals. p91 and p84 indicate Stat 1 a- and b-splice
`variants, respectively, based on their molecular weight. Data are from representative experiments in (A) QGP1 and (B) BON cells.
`
`Roxane Labs., Inc.
`Exhibit 1039
`Page 005
`
`

`
`740 DETJEN ET AL.
`
`GASTROENTEROLOGY Vol. 118, No. 4
`
`After activation of latent Stat transcription factors by
`phosphorylation, Stat1 and Stat2 form heterodimers that
`translocate into the nucleus, where they initiate transcrip-
`tion from ISREs. To directly determine IFN-a–depen-
`dent transactivation of ISREs in NE cells, reporter gene
`studies were performed. In transient transfection experi-
`ments using BON cells, IFN-a dose-dependently stimu-
`lated transcription from an ISRE-luciferase reporter
`construct, but not from a mutant ISRE (Figure 4A),
`indicating that IFN-a–induced Stat signaling results in
`specific transactivation of IFN-regulated genes in this
`cell line. Because we repeatedly failed to achieve accept-
`able expression levels of the reporter construct in trans-
`fected QGP1 cells, nuclear translocation was determined
`as a substitute for reporter gene studies. IFN-a stimula-
`tion led to a time-dependent increase of Stat1 and Stat2
`proteins in nuclear extracts of QGP1 cells (Figure 4B),
`documenting the translocation of activated Stat proteins
`in response to IFN-a.
`IFN-a Inhibits Growth of Human NE
`Tumor Cells
`Having established an intact IFN-a–signaling
`machinery in both cell lines, we next addressed the
`question of whether IFN-a exhibits a direct antiprolifera-
`tive action on QGP1 and BON cells. Growth curves in
`the presence or absence of IFN-a were obtained over a
`period of 6 days (Figure 5A). IFN-a treatment (1000
`IU/mL) resulted in a significant inhibition of prolifera-
`tion in QGP1 (Figure 5A) and in BON cells (Figure 5B).
`The growth-inhibitory effects of IFN-a were concentra-
`tion dependent (Figure 5C and D) with half maximal
`effects on proliferation occurring at less than 100 IU/mL
`in QGP1 and at approximately 500 IU/mL in BON cells,
`which is within the therapeutically achievable range of
`plasma levels in humans.
`
`Growth of malignant tumors in vivo is influenced by
`unique properties of transformed cells, which are more
`closely reflected by anchorage-independent growth assays
`than by substrate-dependent proliferation assays. To
`establish that IFN-a is capable of directly regulating
`anchorage-independent growth of NE tumor cells, colony
`formation of QGP1 and BON cells in agar suspension
`was evaluated. IFN-a treatment profoundly inhibited
`substrate-independent growth of QGP1 cells in a dose-
`dependent manner (Figure 6A) with a maximal decrease
`in colony number to 15% 6 5% of control achieved with
`1000 IU/mL IFN-a. Similarly, substrate-independent
`growth of BON cells was impaired (Figure 6B), although
`the maximal reduction to 77% of control was only
`moderate and did not reach statistical significance. Thus,
`both substrate-dependent and substrate-independent
`growth of NE tumor cells were directly inhibited by
`IFN-a.
`IFN-a–Treated NE Tumor Cells Accumulate
`in the S Phase of the Cell Cycle
`Depending on the cell model investigated, various
`mechanisms responsible for the growth-inhibitory effects
`of IFN-a have been suggested, most of which involve
`modulation of Rb. To test whether Rb might also be
`involved in IFN-a–mediated growth inhibition of NE
`tumor cells,
`immunoblotting experiments were per-
`formed to analyze expression and phosphorylation status
`of Rb in QGP1 and BON cells (Figure 7). These blots
`revealed a striking difference in the Rb content of the 2
`cell lines. Bands representing the hypophosphorylated
`and hyperphosphorylated forms of Rb were easily de-
`tected both in BON and exocrine Capan1 cells, but no
`signal could be visualized in QGP1 cells when equivalent
`amounts of protein were examined. Control blots using
`an antibody against p16ink4a showed a complementary
`
`Figure 4. IFN-a stimulation results in transactivation of an ISRE-driven reporter construct. (A) Dose-dependent increase of relative luciferase
`activity (RLU) in IFN-a–treated BON cells transiently transfected with pGL2-ISRE-luc, but not mutant pGL2-DISRE-luc. Data represent means 6 SEM
`from at least 3 separate experiments conducted in triplicate. (B) Immunoblot analysis of Stat1 and Stat2 in nuclear extracts of QGP1 cells,
`confirming the nuclear translocation of Stat transcription factors in cells stimulated with 1000 IU/mL IFN-a for the indicated times.
`
`Roxane Labs., Inc.
`Exhibit 1039
`Page 006
`
`

`
`April 2000
`
`IFN–a IN NEUROENDOCRINE TUMOR CELLS 741
`
`Figure 5. IFN-a inhibits growth of NE tumor cells in a time- and dose-dependent manner. Subconfluent cells were treated with (A and B) 1000
`IU/mL IFN-a (d) or vehicle (s) for the indicated time periods or with (Cand D) 100 (h), 500 (j), and 1000 (D) IU/mL for the indicated times.
`Viable cells were then manually counted in a hemacytometer. Data represent means 6 SEM from at least 3–4 separate experiments, each
`conducted in triplicate. *P# 0.05 compared with vehicle-treated controls.
`
`expression pattern, as previously described for a variety of
`cells.30 Neither the expression level of p16ink4a nor the
`relative abundance of the hypophosphorylated form of Rb
`were up-regulated by IFN-a treatment, whereas overall
`pRb content was even moderately reduced at 48 hours
`(Figure 7), indicating that the growth inhibition ob-
`served in NE tumor cells did not require activation of Rb.
`To identify potential mechanisms responsible for the
`growth-regulatory effects of IFN-a in NE cells, IFN-a–
`induced cell cycle redistribution was determined by flow
`
`cytometry (Figure 8). In these experiments on unsynchro-
`nized cells, IFN-a treatment increased the fraction of
`cells in the S phase of the cell cycle, whereas the
`proportion of cells in the G1 phase decreased (Figure 8A).
`Both QGP1 and BON cell lines displayed a comparable
`redistribution pattern (Figure 8B), again supporting the
`notion that IFN-a elicits Rb-independent cell cycle–
`specific effects in NE cells. Of note, no population with
`
`Figure 6. IFN-a inhibits anchorage-independent growth of NE tumor
`cells. A single cell agar supension of (A) QGP1 or (B) BON cells was
`incubated with the indicated doses of IFN-a for 10 days, and colony
`formation was evaluated. Data represent means 6 SEM from at least
`3 separate experiments, each conducted in triplicate. *P# 0.05 and
`**P# 0.01 compared with vehicle-treated controls.
`
`Figure 7. IFN-a treatment does not affect Rb phosphorylation in NE
`tumor cells. (A) Immunoblot analysis of Rb expression and phosphory-
`lation status in control and IFN-a–stimulated NE tumor cells as well as
`Capan1 exocrine pancreatic cancer cells. In each lane, 15 µg of whole
`cell lysates was separated by 7.5% SDS-PAGE. (B) Aliquots of the
`same lysates immunoblotted for expression of the CKI p16 ink4a.
`
`Roxane Labs., Inc.
`Exhibit 1039
`Page 007
`
`

`
`742 DETJEN ET AL.
`
`GASTROENTEROLOGY Vol. 118, No. 4
`
`Figure 8. IFN-a–treated NE tumor cells accumulate in the S phase of
`the cell cycle. (A) Representative fluorescence-activated cell sorter
`analysis showing changes in cell cycle distribution of QGP1 cells
`treated with 1000 IU/mL IFN-a for 48 hours. (B) Summary of
`IFN-a–induced cell cycle redistribution in QGP1 and BON cells. Data
`are expressed as mean 6 SEM of the percentage of cells in the S
`phase (closedsymbols) and the G1 phase (opensymbols) compared
`with untreated control cultures, determined in 3 independent experi-
`ments. *P# 0.05 and **P# 0.01 compared with control.
`
`subdiploid DNA content was observed in these experi-
`ments, indicating that no relevant induction of apoptotic
`cell death had occurred. Analysis of the kinetics of
`IFN-a–induced cell cycle effects revealed first S-pha